Total Synthesis of the Glycosylated Macrolide Antibiotic Fidaxomicin

Jun 30, 2015 - The first enantioselective total synthesis of fidaxomicin, also known as tiacumicin B or lipiarmycin A3, is reported. This novel glycos...
0 downloads 7 Views 483KB Size
This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Letter pubs.acs.org/OrgLett

Total Synthesis of the Glycosylated Macrolide Antibiotic Fidaxomicin Elias Kaufmann, Hiromu Hattori, Hideki Miyatake-Ondozabal, and Karl Gademann* Department of Chemistry, University of Basel, St. Johanns-Ring 19, 4056 Basel, Switzerland S Supporting Information *

ABSTRACT: The first enantioselective total synthesis of fidaxomicin, also known as tiacumicin B or lipiarmycin A3, is reported. This novel glycosylated macrolide antibiotic is used in the clinic for the treatment of Clostridium dif f icile infections. Key features of the synthesis involve a rapid and high-yielding access to the noviose, rhamnose, and orsellinic acid precursors; the first example of a β-selective noviosylation; an effective Suzuki coupling of highly functionalized substrates; and a ring-closing metathesis reaction of a noviosylated dienoate precursor. Careful selection of protecting groups allowed for a complete deprotection yielding totally synthetic fidaxomicin. diol β-mannose problem);14 and (3) a suitable protecting group strategy allowing for selective deprotection. In this study, we report on the first total synthesis of fidaxomicin (1). Our highly convergent synthetic route assembles five main fragments by a β-selective noviosylation, a Suzuki cross-coupling, a ring-closing metathesis (RCM), and a β-rhamnosylation.

A

ntibacterial resistance constitutes a growing concern in public health for many societies all over the world,1 and many antibiotics have become less effective if not ineffective against many pathogens.2 This is particularly striking given that the last discovery of a new class of antibiotics dates back to the 1980s. Tuberculosis (TB) has become a threatening example among the diseases that evolved to be resistant to common antibiotic therapy. In 2012 alone, out of 8.7 million new cases of TB, 450 000 patients were infected by multidrug resistantTB.1 With increased frequency, frontline antibiotics (such as rifampicin or isoniazid) fail in therapeutic treatment. All these issues call for new antibiotics to fill this gap.3 To this end, we became interested in fidaxomicin (1, tiacumicin B, lipiarmycin A3),4 which inhibits RNA polymerase in bacteria and shows high activity against drug resistant strains of Mycobacterium tuberculosis.5 Fidaxomicin (1) was FDA-approved in 2011 for the treatment of Clostridium dif f icile infections involved in nosocomial (hospital acquired) diarrhea. Even though advantageous for the therapy of CDI, the low bioavailability of fidaxomicin (1) in the plasma by oral application prohibits its use as a therapeutic against systemic diseases such as TB. In addition to the interesting bioactivity, the 18-membered macrolactone constitutes an architecturally complex synthetic target featuring multiple stereogenic centers and a high degree of unsaturation. The central macrolide is β-linked to a unique D-noviose and D-rhamose, which is complemented with a dichloro homoorselinic acid unit. Despite the fact that fidaxomicin (1) has been isolated in the 1970s and is clinically used to date, no total synthesis of a member of this class has been reported in the literature.7 Coincidentally, syntheses of the core aglycon have been recently published back-to-back-to-back by Altmann et al.,8 Zhu et al.,9 and our group.10 Furthermore, synthetic studies have been performed on 4-OMe D-noviose11 carbohydrates12 and a resorcinol unit.13 The challenges of fidaxomicin (1) concerning its total synthesis reside in (1) rapid access to the carbohydrate and orsellinate building blocks; (2) two challenging β-selective glycosylations (cf. the cis-1,2© XXXX American Chemical Society

The synthesis commenced with the preparation of the carbohydrate and orsellinic acid precursors. For the preparation of the protected resorcylate 4, we chose the biomimetic aromatization strategy established by Barrett et al.15 (Scheme 1). Therefore, the dianion of the known keto dioxinone 216,17 was reacted with freshly prepared propionyl imidazole18 to furnish the corresponding diketo dioxinone. The subsequent aromatization under basic conditions directly yielded the resorcylate 3 (57%). Chlorination with sulfuryl chloride19,20 Scheme 1. Synthesis of the Protected Resorcylate 4

Received: June 1, 2015

A

DOI: 10.1021/acs.orglett.5b01602 Org. Lett. XXXX, XXX, XXX−XXX

Letter

Organic Letters proceeded in quantitative yield, and the allylation gave rise to the protected homoorsellinic acid 4 in excellent yield (95%). The preparation of the rhamnose unit 9 originated in the known D-rhamnoside 521 (Scheme 2), which was synthesized

Scheme 3. Synthesis of Noviosylbromide 16

Scheme 2. Synthesis of the Rhamnosyl Donor 9

at high temperature, first to hydrolyze the acetonide and second to isomerize the furanoside to the pyranoside. As the equilibrium of the 5- vs 6-membered ring was found to be only 1:2 in favor of the desired pyranoside, the furanoside was recovered and resubmitted twice to the same reaction conditions. In this way, the unfunctionalized novioside was obtained as an inseparable anomeric mixture in good overall yield (81%, α/β = 2:1). Based on the previous studies in the area of β-rhamnosylation,14 we opted for introducing the electron-withdrawing 2,3-carbonate protecting group with CDI, which was followed by esterification with isobutyryl chloride to yield the two separable anomers of 15 (67% over two steps). Finally, the acetal 15 (either epimer) was converted to the desired glycosyl bromide donor 16 using HBr in acetic acid. Due to the instability of glycosyl bromides, the intermediate was used in the next step without purification. At this point we aimed for the assembly of the three building blocks 9, 16, and the previously prepared aglycon.10 Even though we achieved satisfactory β-rhamnosylation on the primary alcohol (not shown), the β-noviosylation on the complete macrolide proved exceptionally challenging. Tedious experimentation using different glycosidation methods consistently resulted only in α-selective glycosylation on the protected macrolide. Hypothesizing that the alcohol in the rigid ring system is poorly accessible for an attack from the sterically hindered face of the glycoside (cf. the β-mannose problem),14 we sought to introduce the noviose unit on a linear, more flexible intermediate at an early stage. Therefore, we prepared the alcohol 18 by reductive cleavage of the PNB group of fragment 17 (99%, Scheme 4), which constituted an intermediate in our aglycon synthesis.10 We were pleased to find that the glycosylation with the sterically demanding, yet more flexible secondary alcohol 18, using Helferich’s conditions,30 furnished the noviosylated fragment 19 with a good α/β ratio (63%, α/β = 1:3, 48% of β-anomer isolated). The relative configuration of the β-glycosidic linkage was assigned by NMR studies (NOESY). To the best of our knowledge, this experiment constitutes the first example of a β-selective noviosylation that has been reported in the literature. To assemble the remaining parts of the macrocycle, fragment 22 was prepared from the known precursor 218 by a Yamaguchi esterification (61%) with

from the corresponding methyl-α-D-mannopyrannoside by a Garegg Samuelsson iodination and subsequent Pd-catalyzed hydrogenation.22,23 Butane-2,3-diacetal protection of the trans3,4-hydroxy groups allowed for the selective installation of the methyl group at O-2, and subsequent deprotection furnished the diol 6 in 72% yield over three steps. Next, the thiophenyl group was installed by ZnI2 mediated trans-acetalization with TMSSPh to afford the thioglycoside 7.24 To our delight, the coupling reaction between the resorcylate 4 and the diol 7 proceeded with excellent regioselectivity at O-4 to yield the ester 8 (66%). Interestingly, the O-3 esterified regioisomer was initially formed exclusively and the ester migration occurred with a prolonged reaction time to give rise to the desired product 8. Next, allylation of the phenolic OH group (96%) and hydrolysis of the thioacetal with NBS in acetone/H2O 10:1 was performed with good yields and selectivity (95%, α/β = 5:1). Finally, the lactol was functionalized with N-phenyl trifluoroacetimidoyl chloride to give the rhamnosyl donor 9.25 The switch of the leaving group was crucial in order to achieve good yields as well as anomeric selectivity in the rhamnosylation step (vide inf ra). The known 6-deoxy-6-iodopyranoside 1026 served as starting material for the preparation of the novioside unit (Scheme 3). First, a Vasella ring contraction by a modified procedure27,28 furnished the olefinic furanose 11 (91%). Treatment of lactol 11 with CSA in methanol and 2,2′-dimethoxypropane gave the furanoside 12 without cleavage of the acetonide. Ozonolysis of the terminal olefin using Marshall’s protocol29 in methanolic sodium hydroxide gave direct access to the ester 13 (62%). The major side product in this oxidative cleavage was the corresponding aldehyde (11%), which could be transformed to the desired ester under the same reaction conditions. Next, the gem-dimethyl groups were introduced using MeMgBr to afford the tertiary alcohol 14 in excellent yield (99%). The alcohol 14 was then treated with TFA in methanol B

DOI: 10.1021/acs.orglett.5b01602 Org. Lett. XXXX, XXX, XXX−XXX

Letter

Organic Letters Scheme 4. Completion of the Total Synthesis

the dienoic acid 20.10 The following Suzuki cross-coupling of the boronate 22 and iodide 19 was highly effective and furnished product 23 in good yield. Fortunately, the basic conditions (Pd(PPh3)4, TlOEt in THF−H2O), first used by Altmann and Glaus in their aglycon synthesis,8 did not affect the base labile carbonate and ester moiety, which can be attributed to the very short reaction time (less than 30 min).31 The subsequent ring-closing metathesis reaction by treatment of the linear fragment 23 with the second generation Grubbs catalyst (20 mol %) for 1 h at 100 °C gave the macrolide with an (E/Z) ratio of 2:1 in 75% yield (54% of E-24). The chromatographic separation of the E/Z isomers, allowed for a subsequent recycling of the (Z) isomer. As a result, the overall yield of (E)-24 in this transformation was increased to 63%. Selective deprotection of the primary TBS ether was achieved with trihydrogenfluoride triethylamine in moderate yield (49%). As anticipated, the β-selective installation of the rhamnosyl side chain turned out to be challenging.14 After an extensive screening, we were able to obtain high β-selectivity and a good yield with the trifluoroacetimidate donor 9 in its 4C1 conformation (α/β = 1:4, estimated by 1H NMR analysis of the crude reaction mixture, 62% of 25).25,14e Conveniently, the protection of the O-3 hydroxy group was unnecessary. The final task for the completion of the total synthesis was the deprotection of the TBS, carbonate, and allyl groups. First, the two TBS groups were removed by treatment with trihydrogenfluoride triethylamine at 50 °C (60%). The following carbonate deprotection proved to be troublesome due to competitive isobutyrate ester hydrolysis. Nevertheless, Barton’s base in wet CH2Cl2 gave satisfactory results (good conversion and purity, as judged by NMR analysis). Finally, Pdcatalyzed allyl-deprotection and two-stage purification (preparative TLC followed by HPLC) gave fully synthetic fidaxomicin

(1, 10% over two steps). The observed overall yield for these transformations can be explained by difficulties in the purification of the final product on a small scale. The identity of the synthetic compound was confirmed by coinjection of synthetic and authentic material on reversed-phase HPLC. Interestingly, as the synthetic material contained residual formate from the HPLC purification, the 1H NMR spectrum did at first not fully match with the spectra of the natural product. Again, by mixing equimolar amounts of synthetic and authentic samples, the 1H NMR spectrum clearly confirmed the identity of the synthetic material to authentic fidaxomicin (1). In this letter, we report the first, enantioselective total synthesis of the glycosylated macrolide antibiotic fidaxomicin (1), which has been isolated over 40 years ago and constitutes a clinically used drug. Key features of the synthetic route include (1) a rapid access to the rhamnosyl side chain, (2) the first βselective noviosylation followed by Suzuki cross-coupling and ring-closing metathesis of complex noviosylated precursors, and (3) a β-selective rhamnosylation. In particular, the stereoselective installment of both 1,2-cis diols in the β-linked carbohydrate units constitutes a notable feature of the approach. This successful total synthesis contributes to a detailed understanding of the chemistry of fidaxomicin and paves the way for the generation of analogs addressing some of the shortcoming of the natural product.



ASSOCIATED CONTENT

S Supporting Information *

Experimental procedures, characterization data, NMR spectra and HPLC chromatogram of fidaxomicin (1). The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.orglett.5b01602. C

DOI: 10.1021/acs.orglett.5b01602 Org. Lett. XXXX, XXX, XXX−XXX

Letter

Organic Letters



van der Marel, G. A.; Codée, J. D. C. Chem. Commun. 2012, 48, 2686. (f) Heuckendorff, M.; Pedersen, C. M.; Bols, M. J. Org. Chem. 2012, 77, 5559. (g) Crich, D.; Picione, J. Org. Lett. 2003, 5, 781. For a review on the challenges in β-rhamnosylations, see: (h) El Ashry, H. E.-S. H.; Rashed, N.; Ibrahim, E.-S. I. Tetrahedron 2008, 64, 10631. (i) Nigudkar, S. S.; Demchenko, A. V. Chem. Sci. 2015, 6, 2687. For the investigations of β-rhamnopyranosides using the 2,3-carbonateprotecting group, see: (j) Backinowsky, L. V.; Balan, N. F.; Shashkov, A. S.; Kochetkov, N. K. Carbohydr. Res. 1980, 84, 225. (k) Gorin, P. A. J.; Perlin, A. S. Can. J. Chem. 1961, 39, 2474. (15) (a) Patel, B. H.; Mason, A. M.; Patel, H.; Coombes, R. C.; Ali, S.; Barrett, A. G. M. J. Org. Chem. 2011, 76, 6209. (b) Anderson, K.; Calo, F.; Pfaffeneder, T.; White, A. J. P.; Barrett, A. G. M. Org. Lett. 2011, 13, 5748. (c) Patel, B. H.; Mason, A. M.; Barrett, A. G. M. Org. Lett. 2011, 13, 5156. (16) Navarro, I.; Pöverlein, C.; Schlingmann, G.; Barrett, A. G. M. J. Org. Chem. 2009, 74, 8139. (17) Sakaki, J.; Suzuki, M.; Kobayashi, S.; Sato, M.; Kaneko, C. Chem. Lett. 1990, 901. (18) Staab, H. A. Angew. Chem., Int. Ed. Engl. 1962, 1, 351. (19) Dornhagen, J.; Scharf, H.-D. Tetrahedron 1985, 41, 173. (20) Nicolaou, K. C.; Rodríguez, R. M.; Mitchell, H. J.; Suzuki, H.; Fylaktakidou, K. C.; Baudoin, O.; van Delft, F. L. Chem. - Eur. J. 2000, 6, 3095. (21) Ley, S. V.; Owen, D. R.; Wesson, K. E. J. Chem. Soc., Perkin Trans. 1 1997, 2805. (22) Skaanderup, P. R.; Poulsen, C. S.; Hyldtoft, L.; Jørgensen, M. R.; Madsen, R. Synthesis 2002, 2002, 1721. (23) Zunk, M.; Kiefel, M. J. Tetrahedron Lett. 2011, 52, 1296. (24) Hanessian, S.; Guindon, Y. Carbohydr. Res. 1980, 86, C3. (25) (a) Yu, B.; Tao, H. Tetrahedron Lett. 2001, 42, 2405. (b) Yu, B.; Sun, J. Chem. Commun. 2010, 46, 4668. A review on the recent progress in glycoside bond formation: (c) Zhu, X.; Schmidt, R. R. Angew. Chem., Int. Ed. 2009, 48, 1900. (26) Kumamoto, H.; Deguchi, K.; Wagata, T.; Furuya, Y.; Odanaka, Y.; Kitade, Y.; Tanaka, H. Tetrahedron 2009, 65, 8007. (27) Bernet, B.; Vasella, A. Helv. Chim. Acta 1979, 62, 1990. (28) Kleban, M.; Kautz, U.; Greul, J.; Hilgers, P.; Kugler, R.; Dong, H.-Q.; Jäger, V. Synthesis 2000, 2000, 1027. (29) Marshall, J. A.; Garofalo, A. W.; Sedrani, R. C. Synlett 1992, 1992, 16. (30) (a) Helferich, B.; Wedemeyer, K. F. Liebigs Ann. Chem. 1949, 563, 139. (b) Schroeder, L. R.; Green, J. W. J. Chem. Soc. C 1966, 530. (c) Chen, Y.; Heeg, M. J.; Braunschweiger, P. G.; Xie, W.; Wang, P. G. Angew. Chem., Int. Ed. 1999, 38, 1768. (31) Frank, S. A.; Chen, H.; Kunz, R. K.; Schnaderbeck, A. M. J.; Roush, W. R. Org. Lett. 2000, 2, 2691.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge partial financial support by the NCCR Molecular Systems Engineering, the Latsis Prize (to K.G.), and the Novartis Early Career Award (to K.G.). We thank Dominik Lotter and Reto Witzig at University of Basel for skillful technical support and Priv.-Doz. Dr. D. Häussinger (University of Basel) for assistance with NMR spectroscopy.



REFERENCES

(1) World Health Organization, Antimicrobial Resistance: Global Report on Surveillance; World Health Organization 2014. (2) For a review about antibacterial resistance worldwide, see: Levy, S. B.; Marshall, B. Nat. Med. 2004, 10, S122. (3) Reviews: (a) Wright, P. M.; Seiple, I. B.; Myers, A. G. Angew. Chem., Int. Ed. 2014, 53, 8840. (b) Kirst, H. A. Expert Opin. Drug Discovery 2013, 8, 479. (4) A review on the lipiarmycins/tiacumicins: Erb, W.; Zhu, J. Nat. Prod. Rep. 2013, 30, 161. For the identity of lipiarmycin A3 and tiacumicin B, see: Bedeschi, A.; Fonte, P.; Fronza, G.; Fuganti, C.; Serra, S. Nat. Prod. Commun. 2014, 9, 237. (5) (a) Kurabachew, M.; Lu, S. H. J.; Krastel, P.; Schmitt, E. K.; Suresh, B. L.; Goh, A.; Knox, J. E.; Ma, N. L.; Jiricek, J.; Beer, D.; Cynamon, M.; Petersen, F.; Dartois, V.; Keller, T.; Dick, T.; Sambandamurthy, V. K. J. Antimicrob. Chemother. 2008, 62, 713. (b) Artsimovitch, I.; Seddon, J.; Sears, P. Clin. Infect. Dis. 2012, 55, S127. (c) Goldstein, E. J. C.; Babakhani, F.; Citron, D. M. Clin. Infect. Dis. 2012, 55, S143. (d) Tupin, A.; Gualtieri, M.; Leonetti, J.-P.; Brodolin, K. EMBO J. 2010, 29, 2527. (e) Duggan, S. T. Drugs 2011, 71, 2445. (6) Sears, P.; Crook, D. W.; Louie, T. J.; Miller, M. A.; Weiss, K. Clin. Infect. Dis. 2012, 55, S116. (7) For isolation and biosynthesis studies, see, for example: (a) Parenti, F.; Pagani, H.; Beretta, G. J. Antibiot. 1975, 28, 247. (b) Coronelli, C.; White, J. R.; Lancini, C. G.; Parenti, F. J. Antibiot. 1975, 28, 253. (c) Coronelli, C.; Parenti, F.; White, R.; Pagani, H. GB 1458512 1973. (d) Xiao, Y.; Li, S.; Niu, S.; Ma, L.; Zhang, G.; Zhang, H.; Zhang, G.; Ju, J.; Zhang, C. J. Am. Chem. Soc. 2011, 133, 1092. (e) Niu, S.; Hu, T.; Li, S.; Xiao, Y.; Ma, L.; Zhang, G.; Zhang, H.; Yang, X.; Ju, J.; Zhang, C. ChemBioChem 2011, 12, 1740. (8) Glaus, F.; Altmann, K.-H. Angew. Chem., Int. Ed. 2015, 54, 1937. (9) Erb, W.; Grassot, J.-M.; Linder, D.; Neuville, L.; Zhu, J. Angew. Chem., Int. Ed. 2015, 54, 1929. (10) Miyatake-Ondozabal, H.; Kaufmann, E.; Gademann, K. Angew. Chem., Int. Ed. 2015, 54, 1933. (11) For studies on the synthesis of 4-OMe D-novioses, see: (a) Pankau, W. M.; Kreiser, W. Tetrahedron Lett. 1998, 39, 2089. (b) Pankau, W. M.; Kreiser, W. Helv. Chim. Acta 1998, 81, 1997. (c) Rajesh, B. M.; Shinde, M. V.; Kannan, M.; Srinivas, G.; Iqbal, J.; Reddy, D. S. RSC Adv. 2013, 3, 20291. (d) Yu, X. M.; Shen, G.; Blagg, B. S. J. J. Org. Chem. 2004, 69, 7375. (e) Reddy, D. S.; Srinivas, G.; Rajesh, B. M.; Kannan, M.; Rajale, T. V.; Iqbal, J. Tetrahedron Lett. 2006, 47, 6373. (12) For studies on the synthesis of 2-OMe rhamnose, see: Lipták, A. Carbohydr. Res. 1982, 107, 300. (13) Alexy, M.; Scharf, H.-D. Liebigs Ann. Chem. 1991, 1991, 1363. (14) (a) Crich, D.; Sun, S. J. Am. Chem. Soc. 1998, 120, 435. (b) Crich, D.; Vinod, A. U.; Picione, J. J. Org. Chem. 2003, 68, 8453. (c) Crich, D.; Sun, S. J. Org. Chem. 1997, 62, 1198. (d) Crich, D.; Chandrasekera, N. S. Angew. Chem., Int. Ed. 2004, 43, 5386. (e) Christina, A. E.; van der Es, D.; Dinkelaar, J.; Overkleeft, H. S.; D

DOI: 10.1021/acs.orglett.5b01602 Org. Lett. XXXX, XXX, XXX−XXX